An Empirical Validation of Bipolar Risk Perception Scaling Methods1

Abstract

Bipolar rating scales (e. g., semantic differentials) are a common tool for gathering perceptions about attributes of persons, events, or activities. Researchers have questioned, however, whether individual responses are methodologically constrained when bipolar scales are used. Comparing both bipolar and unipolar scales, the present study investigates previous research findings on perceptions of risk by Fischhoff, Slovic, Lichtenstein, Read, and Combs (1978). The subjects in the current study (N= 94) rated the same 30 technologies, substances, and activities as in the Fischhoff et al. (1978) study but used unipolar scales. Correlational analysis indicated that some presumed antagonistic adjective pairs were indeed opposites, while others were not. On the other hand, factor analyses indicated that the same factors described by Fischhoff et al. (1978) were uncovered in the present study. However, the use of a unipolar methodology allowed us to refine the interpretations of both factors.